How Long Should a Short-Form Video Be?

There is no magic number — length only matters through retention. Ranges that work by content type, how to tell when a clip is too long, and how to cut a great moment down without losing it.

Clipo AI September 17, 2026 9 min read

A short-form video timeline trimmed to its strongest section

“How long should my Short be?” is usually asked as if there were a number the algorithm is secretly looking for. There isn't. Platforms don't reward length; they reward people choosing to keep watching. Length only matters because it changes how hard that is.

That reframing is more useful than any target, because it tells you what to do with a clip that's too long — and it's almost never “cut it to 30 seconds.”

Length is a proxy. Retention is the real metric

Two numbers get confused here: average view duration (how many seconds people watch) and percentage viewed (how much of the video that is). A 20-second clip watched to 90% and a 60-second clip watched to 40% can have similar average durations, and they send very different signals.

The first says “people who start this finish it.” The second says “most people leave before the point.” Short-form distribution leans heavily on the first kind of signal — which is why a tight clip often outperforms a longer one that contains more good material.

So the useful question isn't “how long?” but “how long can this particular idea hold attention?” A sharp opinion runs out of road fast. A good story can hold a minute and a half without effort.

The limits are ceilings, not targets

The major platforms now accept short-form uploads well past a minute — YouTube Shorts and Instagram Reels both allow up to about three minutes, and TikTok far longer. These limits change often, so check the current numbers before you plan around them.

What they don't mean is that three minutes is a good length. A ceiling tells you what's allowed, not what works. Most of the clips that travel are well under the limit, because most ideas don't need the room.

Ranges that work, by what the clip is doing

Rather than one number, match the length to the job. These are starting points to test against, not rules.

One idea or a hot take: 15–30 seconds

A single claim, stated and justified. Anything past 30 seconds is usually the speaker repeating themselves. These are the clips most likely to be rewatched, because they're over before the viewer has decided to leave.

An explainer or tip: 30–60 seconds

One concept, explained properly. The sweet spot for most educational content. If it takes longer than a minute, it's often two tips that should be two clips.

A story or interview moment: 45–90 seconds

Stories earn length because the viewer wants to know how they end. The setup can breathe a little — but only if the payoff is genuinely worth it, and only if the first sentence already promises it.

Past 90 seconds: only with a late payoff

Longer clips work when the reason to watch arrives late and is worth waiting for — a reveal, a result, a turn. If the best line is in the first 20 seconds, the rest of a two-minute clip is just places for people to leave.

How to tell a clip is too long

You don't need to guess. The retention graph shows you exactly where attention goes.

A slope that steepens mid-clip — people were fine, then something lost them. Look at what's on screen at that timestamp
A dead middle — the clip has a good start and a good end with a stretch in between where nothing new is said
Setup that could be a sentence — if 15 seconds of context could be replaced by one line of on-screen text, it should be
An ending that trails off — the point landed, then the speaker kept going. Cut on the point, not after it
A second idea — if the clip changes topic partway, you have two clips glued together

Cutting a great moment down without losing it

The hard case is a clip that's too long but genuinely good throughout. Four edits handle most of it, in this order:

Start at the conclusion. Move the sharpest sentence to the front, then let the reasoning follow
Remove the throat-clearing. “So, what I'd say is…” and every restart before the real sentence
Tighten the pauses. Dead air between sentences adds up faster than anything else — often 10 seconds in a minute
Split, don't squeeze. If it's still too long, find the natural break and make it two posts. Part one's ending becomes part two's hook

Loops and rewatches

Short clips have an advantage long ones can't match: a viewer can watch them twice without deciding to. A clip whose last line leads naturally back into its first gets replayed, and those replays count.

You can build for this deliberately — end on a question the opening answers, or cut so the final frame matches the first. It only works on short clips, which is one more reason not to pad.

A simple way to settle it for your audience

Published advice can't tell you what your specific audience tolerates. Your own numbers can. Take one strong moment and publish two cuts a week or two apart — a tight version and a fuller one — and compare percentage viewed and shares, not views alone.

Do that a handful of times and you'll know your own ranges far better than any chart can tell you.

Pace changes the length that works

Two clips of identical length can feel completely different. A 60-second clip with a cut, a caption change or a new visual every few seconds feels quick; the same 60 seconds as one static shot of someone talking feels long, even when what they're saying is good.

That's why length advice without pace advice is misleading. Before cutting a clip shorter, try making it move: tighten the pauses, punch in on alternate sentences, and let the captions carry emphasis. A clip that feels long at 45 seconds often feels right at the same length once it has rhythm.

The reverse is also true. Speeding up a clip that has nothing new to say doesn't fix it — it just gets to the dead middle faster. Pace can make a good clip feel shorter; it can't make a padded one worth watching.

Key takeaways

Platforms reward people choosing to keep watching — length only matters through that.
Upload limits are ceilings, not targets; most clips that travel are well under them.
Match length to the job: a take is 15–30s, an explainer 30–60s, a story up to 90s.
The retention graph tells you where a clip is too long — cut there.
If a clip is good but long, split it rather than squeezing it.

Frequently asked questions

What is the best length for a YouTube Short?

There isn't one number. Most Shorts that perform well sit between 15 and 60 seconds, because most single ideas fit in that range. Longer Shorts work when the payoff arrives late and is worth the wait.

Do longer videos get more reach?

Not by being longer. A longer clip has more total watch time available, but only if people actually stay — and a lower percentage viewed usually costs more distribution than the extra seconds earn.

Should I cut every clip to under 30 seconds?

No. Cutting a story to 30 seconds usually removes the part that made it worth telling. Cut to the length the idea needs, then remove everything that isn't the idea.

How do I know where people stop watching?

Every major platform shows a retention graph for each video. Look for where the line drops faster than elsewhere, then look at what's on screen at that timestamp — it's almost always a slow setup, a repeated point or a change of topic.

Is it better to post one long clip or two short ones?

If the material contains two distinct ideas, two clips nearly always do better — each gets its own hook, and the first can send viewers to the second. One idea stretched across two clips is worse than either.

Does length matter differently on LinkedIn?

Slightly. LinkedIn audiences tolerate a slower start and reward substance, so 30–90 seconds is comfortable for a clear point. The same rule still applies, though: stop when the idea stops.

Keep reading

Find the strongest 30 seconds automatically

Clipo scans a long recording for its best moments and cuts them to a length that holds attention — reframed, captioned and ready to post.

Try Clipo Free